Browsing Tag

AI governance

22 posts

Abu Dhabi Wants an AI-Native Government. What Happens to Public Work When the Rules Become Machine-Readable?

A bright, empty modern government service hall with white counters, glass walls and morning sunlight over a desert city
On 2 November 2026 Abu Dhabi signed four agreements with OpenAI: ADNOC, Mubadala, a five-year rollout with G42’s Inception42 and a “Machine-Readable Government” pilot that lets AI assist officials with routine compliance checks. What was signed, what was not disclosed, and what it means for companies, residents and public employees when the rulebook becomes software.
View Post

Rules That Keep Humans in Charge: What Microsoft’s School AI Pact and California’s New Laws Actually Say

A bright, empty school classroom in morning light with rows of wooden desks, closed laptops and a blank whiteboard
Microsoft now offers every US school district a contract that bans training on student data, tracking, AI companions and AI decisions without human review. California has banned AI that reads workers’ emotions and firing by algorithm alone, and New York has restricted chatbot companions for children. What the texts actually say, where they are narrower than the headlines, and what it means for companies, schools, workers and parents.
View Post

UBS Staff Now Run 12,700 Personal AI Agents. What Agent Work Looks Like Inside a Bank, and What UBS Doesn’t Say

A bright open-plan office with rows of empty white desks, each with a small glowing glass cube, and a lake with mountains behind the windows
UBS says its employees run more than 12,700 personal AI agents, next to 608 live AI use cases and 18,000 developers using AI tools. What agent work looks like inside a regulated bank, why the 1,500-job drop in the same quarter is the Credit Suisse integration by UBS’s account, what UBS does not disclose about governance, and what it means for companies and employees.
View Post

OpenAI, Google, Microsoft and Anthropic Back One Opt-Out From AI Training. What It Covers, What It Doesn’t, How to Set It

A blank white paper note clipped to a bright blue front door of a white townhouse, with a potted plant beside the door
OpenAI, Google, Microsoft and Anthropic say they will honor one shared robots.txt rule against AI training, from an IETF draft published on 30 October. Anthropic says it has started, OpenAI aims for the coming months, Microsoft starts on 1 February 2027 and Google in 2027. What the signal covers, what it does not (AI answers, crawling, payment), whether it is binding, and how site operators set it.
View Post

Who Tests the Most Powerful AI Before It Ships? Britain Pays Far More Than the US, but Not “10 Times” More

A sealed glass test chamber on a white laboratory bench with a softly glowing turquoise cube inside and two pressure gauges
Britain just gave its AI Security Institute an extra £120 million, spread over four years. The US testing center at NIST runs on about $15 million a year. Per year, Britain spends roughly 8 to 9 times more, not “10 times” from the top-up alone. What the testers found in 2026, why access matters more than money, how the EU AI Office differs, and what it means for companies and users.
View Post

AI Agents Now Start 4 in 10 Pull Requests on GitHub. The Bottleneck Has Moved to Review

A long line of white paper folders on a conveyor belt leading to a small wooden desk with a brass lamp and a magnifying glass
GitHub’s Octoverse 2026 says more than 40% of pull requests in September started with an AI agent’s first draft, up from 17% in March. What GitHub actually counted, why a 90% merge rate is not a quality score, what Copilot code review Max, approvals and Fusion cost, who answers for agent code that ships, and what it means for companies and developers.
View Post

Did Anthropic Train Claude on Pirated Books? What Is Alleged, What Is Proven, and What It Means for the IPO

A tall wooden library bookshelf with turquoise book spines, thin streams of light flowing from the books toward a small computer server in a bright room
The Stanford Daily says Anthropic illegally trained Claude on professors’ books. The court record says something narrower: training on books was fair use, downloading over seven million pirated copies was not, and the case settled for $1.5 billion before trial. Here is what is alleged, what is proven, what the S-1 really says about blackmail and sentience, and what it means for companies, authors, users and investors.
View Post